Charting Transitions to Conservation-Friendly AgricultureThe RainfoResT alliance’s appRoach To moniToRing and assessing

ResulTs foR biodiveRsiTy, ecosysTems and The enviRonmenT

charting Transitions to conservation-friendly agriculture

performance measurement 101

community mapping

The natural ecosystem assessment

Water Quality monitoring

carbon and greenhouse gases

Reducing pollution and sustaining on-farm natural Resources

community-based species monitoring

next steps for charting Transitions to conservation-friendly agriculture

3

4

7

8

11

13

13

15

15

December 2013

The Rainforest Alliance works

to conserve biodiversity and

ensure sustain-able livelihoods by transforming

land-use prac-tices, business practices and

consumer behavior.

www.rainforest -alliance.org

contents

Report by the Rainforest

AllianceEvaluation

& Research Program

Contributors:

Jeffrey C. MilderDavid Hughell

Deanna NewsomWilliam CrosseCrispen Wilson

Elizabeth T. Kennedy

This work has been made possible by

the generous financial support

of the Z Zurich Foundation

(“Foundation”), a private founda-tion funded by

Zurich Insurance Company Ltd

and Zurich Life Insurance

Company Ltd (together

“Zurich”). The content of

this publica-tion reflects

the opinion of the Rainforest

Alliance and not necessar-

ily that of the Foundation

or Zurich, and neither can be

held liable in this regard.

This document has been produced through the support of the Biodiversity and Agricultural Commodities Program (BACP) as part of a project to increase sustainable cocoa production in order to achieve biodiversity conservation outcomes in South Sulawesi, Indonesia. Additional support has been provided by the Z Zurich Foundation.

Cover image: Noah Jackson

3

The need to make agriculture more efficient, less polluting, and less damaging to wildlife and eco-systems is widely acknowledged.

The Rainforest alliance’s pursues these goals on the ground through two linked strategies: the training of farmers and farmer groups on best practices for productive, efficient, and conserva-tion-friendly agriculture; and the certification of farms that have adopted these practices, which are based on the sustainable agriculture net-work sustainability standard. Worldwide, this system has now been applied across 3 million hectares, involving more than 900,000 farmers growing over 50 crops.

certification provides an important independent validation of sustainable practices. but many farmers and buyers of certified crops want to go beyond the certification label to understand

the real effects of these changes on the ground. They want to be able to document and claim success for positive results, understand the rea-sons behind challenges and setbacks, and learn how they can improve results over time. other stakeholders—from rural communities to gov-ernment regulators—want to know how certi-fication contributes to important goals such as protecting watersheds and conserving wildlife.

To meet these needs, the Rainforest alliance has developed a robust system and set of tools to monitor environmental performance and track results over time. below, we explain the envi-ronmental performance monitoring system and introduce its component tools to explain how the Rainforest alliance is charting transitions to conservation-friendly agriculture to meet the in-formation needs of farmers, buyers and other stakeholders.

FIGURE 1: The Sustainable Agriculture Network (SAN) agriculture standard specifies principles and criteria for socially and environmental-ly responsible crop produc-tion. The colored boxes to the right highlight some of the SAN criteria that most directly deliver environmental benefits and mitigate risks.

Charting Transitions to Conservation-Friendly AgricultureThe RainfoResT alliance’s appRoach To moniToRing and assessing ResulTs foR biodiveRsiTy, ecosysTems and The enviRonmenT

Monitoring is not an academic exercise, but rather a way to answer questions about wheth-er the SAN standard and the Rainforest Alli-ance’s training programs are delivering the desired benefits for farmers, communities and society at large. With this in mind, our monitor-ing approach is guided by the specific interests and needs of stakeholders in understanding re-sults in any given landscape or supply chain. It is also guided by our “theory of change,” a logi-cal model of how certification and training help protect the environment. Figure 1 illustrates some of the key potential environmental ben-efits, including improved on-farm conservation value, maintenance of key ecosystem services, reduction in pollution and other negative exter-nalities, reduction in hunting and other direct threats to biodiversity, and contribution to ecological integrity beyond farm boundaries.

4

Performance Measurement 101: Deciding What to Monitor, and Why

FIGURE 2: A clear theory

of change drives the Rainforest

Alliance’s work to ensure that field activities

directly support key desired

outcomes and to help define

monitoring and evaluation

needs. The diagram

illustrates a portion of the

agriculture theory of

change focused on

environmental results.

WILLIAM CRoSSE

5

Starting from this framework, monitoring in any given landscape or supply chain will typi-cally be oriented around a subset of results areas that are most relevant to that specific context or of greatest concern to stakeholders. For instance, where farming communities abut a protected area, it may be important to moni-tor encroachment and understand changes in landscape connectivity. Where the landscape is dominated by agriculture, farmers and com-modity buyers may be most interested to un-derstand the sustainability of key natural assets that sustain productivity, such as water, soils and on-farm vegetation. In this way, monitoring and assessment becomes a demand-driven and hypothesis-driven activity that answers ques-tions about effectiveness and points the way to future adaptations that can improve outcomes.

As shown in Figure 2, certification and training can deliver results at multiple levels. A direct result may be the adoption of “better manage-ment practices” (BMPs), such as reduced pes-ticide use, increased soil surface cover or im-proved shade management for shade-tolerant

perennial crops. These practices, in turn, may result in changes in environmental conditions either at the farm scale (e.g., reduced erosion or improved riparian habitat) or across larger areas and longer time frames (e.g., improved habitat connectivity or water quality at a land-scape or catchment scale). For some environ-mental results, such as the reduction in toxic pesticide pollution, it is most feasible and suf-ficiently accurate to consider BMP adoption as a proxy for the delivery of benefits. For others results, where linkages between BMPs and outcomes are more complex, it is neces-sary to measure actual changes in ecosystems. In some cases, we track change at the level of both practices and outcomes, in order to un-derstand the degree to which BMPs translate into on-the-ground environmental sustainabil-ity, and which factors might support or impede this process.

With these considerations in mind, the Rain-forest Alliance has developed a suite of field tools and methods to monitor BMPs and/or outcomes across key environmental results ar-

Students interviewing a farmer in Bantaeng, Indonesia, to take stock of existing farm-ing practices and identify opportunities for improve-ment

WILLIAM CRoSSE

6

eas, including natural ecosystems, water, pol-lution, on-farm natural assets (a basis for key ecosystem services), greenhouse gases and species of conservation concern. These meth-ods are summarized in Table 1 and explained in more detail below. At the end of this brief, we provide an example of how these tools are combined—alongside monitoring of productiv-ity, social and economic outcomes—to assess the sustainability of farming systems over time.

ASSESSMENT METHoD WHAT, WHy, AND HoW WE APPLy THE METHoD

context mapping (including community mapping)

natural ecosystem assessment

Water quality monitoring

farm performance monitoring Tool

carbon and greenhouse gases

community-based species monitoring

mapping processes that combine local knowledge with other geo-graphic data to develop landscape maps identifying key conserva-tion threats and opportunities, as well as farm boundaries, that enable spatially explicit analysis of landscape change

a suite of three different methods used alone or in combination to track changes in: a) on-farm vegetation, including tree diversity and structure; b) land use on and adjacent to certified farms; and c) broader effects on forest encroachment, conserva-tion, and connectivity

cost-effective field protocols to monitor key stream health indicators as proxies for eutrophication, sedimentation, pesti-cide pollution, and aquatic biodiversity

method to track changes in agronomic and environmental bmps based on farmer interviews and direct observation; provides de-tailed information on direct effects of training and certification on key areas, such as soil conservation and fertility management, agrochemical use, waste management, and tree planting

method that adapts and augments the cool farm Tool to estimate a farm’s greenhouse gas “footprint,” as well as carbon sequestra-tion associated with the conservation of non-cropped vegetation on the farm

participatory monitoring of wildlife by farmers and community members, based on photos of locally occurring species of conser-vation concern

All of these monitoring and assessment meth-ods, except for the initial context mapping, are carried out as time-series investigations to evaluate trajectories of change. Typically, a baseline assessment will be conducted prior to training and certification, followed by repeat surveys after the training and certification and in subsequent years. When carried out with a control group (for example, non-certified farms), the methodologies are adequate to at-

TABLE 1:Summary of

ecosystem and environmental

field-assessment

methods

WILLIAM CRoSSE

Measuringcacao pods

7

tribute specific environmental changes to the training and certification activities. In the ab-sence of control groups, performance trajecto-ries can still be interpreted in light of the re-sults chain, but with a lower degree of certainty regarding the attribution of outcomes to the training and certification activities.

These tools are not applied at every site where the Rainforest Alliance conducts train-ing or certification, but rather at a sample of locations where environmental conditions and stakeholder interest create a strong mandate to track and manage conservation results. In the future, a simplified version of some of these tools may be integrated into the certification audit process so that they could be applied across much or all of the Rainforest Alliance’s agriculture portfolio.

Data generated through these methods be-come a valuable asset not only for the Rainfor-est Alliance to evaluate its programs, but also for farmers, commodity buyers and others to understand and better manage production systems and product flows. To this end, data are reported through various summaries and analytics that can be used by farmer groups, training managers, and companies, as well as external stakeholders (e.g., project donors).

To understand the conservation challenges and agricultural context in any given landscape, it is helpful to begin with a bird’s-eye view of the landscape. Typically, this view is provided by a combination of professional mapping, using Geographic Information System (GIS) tools and community mapping carried out through farmer groups and community meetings. GIS-based mapping and spatial analysis begins by

Community Mapping and Spatial Analysis

interpreting aerial imagery to create a land-cover map for the focal area (where training and certification will be carried out) plus a surrounding area that may be considered the

“zone of influence” where certified agriculture interfaces most directly with other land uses, such as non-certified crops, forests and pro-tected areas. The map distinguishes key land uses, such as agroforestry with low and high shade cover, forest, annual crops, residential areas and pasture. The map is then used to answer questions such as:

1. What is the landscape composition? What proportion of the landscape is under conservation friendly management versus other uses?

2. What is the landscape configuration? Where are cropped areas located relative to forests and other natural areas, and what does this imply for habitat connec-tivity? Is the landscape highly fragmented or well-articulated into separate produc-tion and conservation areas?

3. Does the landscape composition and con-figuration pose any specific risks or opportunities for conservation-friendly agriculture? For instance, does the focal area abut protected areas where there may be a risk of encroachment? Does it encompass very steep slopes at risk for landslides?

Landscape composition and configuration met-rics can be quantified, so that baseline analysis may be compared to the condition in subse-quent years, following training and certifica-tion. Figure 3 shows an example of a landscape context map used to understand how cocoa certification may contribute to buffering pro-tected areas and restoring a wildlife corridor in western Ghana.

8

While land-cover maps provide a critical land-scape context to understand transitions to conservation-friendly agriculture, they miss important details that can be understood only by “peering beneath the canopy” and consult-ing local residents. For this reason, we some-times augment the GIS context mapping pro-cess with participatory “community mapping” sessions within the landscape. During these sessions, Rainforest Alliance trainers typically convene groups of a few dozen farmers around a large-format map of the landscape based on high resolution satellite imagery. This activity usually generates great excitement, as it is of-ten the first time many of the farmers have seen their farms from above. The trainers help ori-ent farmers to the map and then invite them to mark up the map to identify points of interest throughout the landscape. These may include the farmers’ individual land holdings, commu-nity forests and sacred areas, water sources, and areas of human conflict with wildlife or other environmental hazards. Community map-ping serves several important purposes. First, it helps orient farmers to the training and cer-tification project. Second, it fosters sensitiv-ity to environmental issues and to the wider landscape as an important management do-main affecting their own livelihood. Once the information on the marked-up maps has been digitized, these maps are returned to the group

to keep, where they often provide an important resource for group planning and decision-mak-ing. Third, community mapping can help define the boundaries of farmers’ land holdings—in-formation that can otherwise be surprisingly elusive in many smallholder-dominated rural areas. Data on farm land areas is used for a range of purposes, from calculating more pre-cise productivity estimates to facilitating farm-ers’ access to government-supplied agricultur-al inputs. Finally, community mapping provides Rainforest Alliance trainers with important in-formation about conservation threats and op-portunities that a project should address.

The Natural Ecosystem Assessment

The purpose of the Natural Ecosystem Assess-ment (NEA) is to assess the condition of natu-ral and semi-natural ecosystems on and near farms working with the Rainforest Alliance, and to evaluate the effects of training and certifica-tion on this condition over time. As has been identified through prior research, efforts to promote conservation-friendly agriculture can benefit biodiversity in three main ways: 1) by increasing the conservation value of the farm itself; 2) by contributing to a “matrix” of land use types that helps protect wildlife outside of protected areas; and 3) by reducing pressure to destroy or degrade remaining forests and other natural ecosystems. The first of two of these represent “land-sharing” approaches that support biodiversity conservation within the farming landscape. The last one is a “land sparing” approach, whereby farming and habi-tat protection are segregated on the landscape, and natural ecosystems are set aside as pro-tected areas.

Rainforest Alliance and the SAN recognize the importance of each approach, and address both in their training and certification activi-ties. However, the types of monitoring needed to evaluate land-sharing approaches differ from those required for land-sparing approaches. And in many settings, the greatest conserva-tion opportunities (and hypothesized benefits of the RA/SAN system) will focus primarily on one or the other conservation approach. For instance, in Ghana’s Bia-Juabeso landscape, a 15 km-wide cocoa-producing landscape sand-wiched between two protected areas, a key goal

FIGURE 3: This map of the

Bia-Juabeso landscape in

western Ghana illustrates the

juxtaposition of cocoa-produc-

ing areas where the Rainforest

Alliance is work-ing with adja-

cent protected areas. The map

is used to define an appropriate

sampling framework for

the Natural Ecosystem As-

sessment, which is used here

to understand impacts of

farmer training and cocoa

certification on reducing forest encroachment and increasing

tree cover in the landscape.

9

is to promote more productive management of agricultural land while stopping encroachment into the critical forest reserves (land-sparing). Conversely, in Bantaeng (South Sulawesi), Indo-nesia, the landscape contains very few remain-ing forest patches yet maintains moderate tree cover and diversity that could be sustained and enhanced through conservation-friendly agro-forestry (land-sharing). With such differences in mind, the NEA is designed as a suite of three core field methods, plus supporting landscape mapping and spatial analysis, so that its appli-cation can be tailored to address the core ele-ments of the biodiversity conservation results chain in any given setting (see Table 2).

In all of the NEA methods, we use vegetation composition and structure as a proxy for con-servation value. Compared to monitoring ani-mal taxa, vegetation surveys have several ad-vantages. First, plants tend to be relatively easy and cost-effective to survey using standard methods, such as vegetation plots. Second, farmers are likely to manage vegetation direct-ly as a result of participation in the RA/SAN sys-tem. Therefore vegetation metrics are likely to

FIGURE 4: Community mapping activities helps introduce farmer groups to conservation-friendly farming practices while allowing local residents to identify key features, challenges, and opportu-nities in the landscape. These may include farm boundaries, conservation areas, wildlife sightings, and areas of envi-ronmental risk such as severe erosion or human-wildlife conflict.

be sensitive and rapidly responding indicators of ecological changes that are directly attribut-able to training and certification. In contrast, assemblages of animals (particularly larger or wider-ranging mammal and bird species) may respond to a variety of factors other than those influenced by the RA/SAN system, or they may respond to change over larger spatial and tem-poral scales influenced by natural population cycles and broader trends. Therefore it may be difficult to interpret the degree to which ob-served changes in animal populations are the results of training and certification, versus oth-er factors. Finally, there is a large body of lit-erature on the relationship between vegetation characteristics and animal diversity covering nearly every part of the tropics. This literature generally provides good guidance for using and interpreting vegetation metrics as proxies for broader biodiversity impacts. Studies on the effects of the RA/SAN system on animal con-servation are an important complement to the vegetation monitoring carried out through the NEA, and they are conducted through more in-depth focused research projects, usually with external research partners.

WILLIAM CRoSSE

10

Table 2:Summary

of the field surveys and analysis in-

cluded in each NEA module. The modules may be used

individually or in combina-

tion at any given site,

depending on the evaluation

questions of interest.

SCALE(S) AT WHICH NATU R AL ECoSySTE M AS S E S S M E NT MoDU LE SMoNIToRING IS CoNDUCTED

landscape scale

farm scale

plot scale (vegetation sampling plots of 10x10m)

landscape change assessment

land cover is classified based on analysis of remote sensing imagery with ground verification

no sampling conducted

if more detailed information is required, then for each land cover type, a sample of plots is surveyed for:• Tree canopy height• Tree canopy % cover• % ground cover• evidence of erosion or other human impacts• species and diameter of non-crop trees

monitoring on-farm conservation value

changes in on-farm values are analyzed relative to surrounding land use and landscape structure

for a representative sample of farms, farm boundaries and land uses are mapped; species and diameter of emergent trees (very large or tall trees) is recorded

for each major land use on the farm, up to three plots are surveyed for:• vegetation struc-ture (% cover at each of six vertical layers)• species and diameter of non-crop trees• characteristics such as density, age, and health of the focal crop (e.g., cocoa)

assessing changes at the forest frontier

changes in condition at the forest frontier are analyzed relative to surrounding land use and landscape structure

no sampling conducted

Ten or more plots are arrayed at regular intervals along a transect spanning the agriculture-forest boundary. each plot is surveyed for:• Tree canopy height• Tree canopy % cover• % ground cover• evidence of erosion or other human impacts• species and diameter of non-crop trees

Field sampling as part of the

Natural Ecosystem

Assessment is important for

understanding how certifica-

tion and training are

affecting on-farm

conservation values such as

diversified, multi-strata native veg-

etation in agroforestry systems and

non-cropped portions of the

farm.

WILLIAM CRoSSE

11

SCALE(S) AT WHICH NATU R AL ECoSySTE M AS S E S S M E NT MoDU LE SMoNIToRING IS CoNDUCTED

landscape scale

farm scale

plot scale (vegetation sampling plots of 10x10m)

Results from the NEA, conducted as a time series before and following training and certification, may be used to answer questions such as:

1. Have training and certification resulted in increased tree canopy density, di-versity, or structural complexity for agroforestry crops? And how does this vary across different types of farms and different parts of the landscape?

2. What has been the effect on emer-gent trees (very tall trees, which have been found to be important habitat features for birds and other wildlife)?

3. How is land cover changing in the land-scape? Are rates of change different on certified farms versus non-certified areas?

4. To what extent is there encroachment into, or degradation of, natural forests in pro-tected areas or immediately adjacent buf-fer areas? Are rates of encroachment dif-ferent where certified farms abut protected areas versus where non-certified land uses abut these areas?

Water Quality Monitoring

Many conventional agriculture practices are known to negatively affect water quality, flow rates, and riparian habitat. The elimination of overstory tree cover and vegetated riparian zones can lead to soil erosion and elevated lev-els of particulate matter in streams, damaging aquatic ecosystems and potentially affecting human health. The spilling or overuse of pes-ticides or other agrochemicals can bring dan-gerous neurotoxins into a watershed, while the overuse of fertilizers can pollute neighboring water bodies, starting a cycle of eutrophication in which oxygen levels drop and fish and other aquatic fauna die. The withdrawal of large vol-umes of water for irrigation, washing or pro-cessing can reduce flow rates, putting aquatic fauna at risk, concentrating toxins and nutri-ents, and leaving insufficient water for down-stream users.

The goal of the Rainforest Alliance’s water qual-ity monitoring methodology is to track the effi-

cacy of water-related BMPs included in the SAN standard in avoiding or ameliorating the above risks. To do this, the methodology includes a suite of four components: 1) a farm assess-ment; 2) probe-based measurement of stan-dard water quality variables; 3) the Streamside Visual Assessment Protocol; and 4) an aquatic invertebrate assessment.

The farm assessment is a survey method used to assess the farm’s water-related management practices, including:

• which water and soil conservation measures are being implemented (e.g. drip irrigation, mulching, rainwater cisterns, low-water cof-fee pulping systems)

• which water contamination prevention mea-sures have been implemented (e.g. cleaning of biocide application equipment at a desig-nated area away from water sources)

• which agrochemicals are used, and at what volume and frequency

• how the farm is disposing of wastewater, including the color and odor of any visible wastewater

Farmer training in Bantaeng, Indonesia

BoNAR MAToNDANG

12

This contextual information is used to identify the primary threats to water quality on a given farm, and to link data on stream water quality and riparian habitat condition to the adoption of certain practices.

To conduct probe-based measurements, we use a multi-parameter probe to measure stan-dard water quality variables such as turbidity, conductivity, flow rates, temperature, pH, am-monia, nitrate, phosphorus, dissolved oxygen and suspended sediment. This choice of pa-rameters closely follows guidance from the U.S. Environmental Protection Agency for the monitoring of surface waters.

The Streamside Visual Assessment Protocol (SVAP) was first developed by the U.S. De-partment of Agriculture to provide an easy, low-tech means to evaluate stream and ripar-ian area condition based on a simple rating scheme. As the name suggests, trained techni-cians use visual observations to rate the fol-lowing elements on a scale from 1 (excellent quality) to 10 (very poor quality):

• water quality (e.g. amount of algae pres-ent, water color)

• physical habitat (e.g., stability of stream banks, amount of canopy cover and downed wood)

• biological condition (e.g. amount of streamside vegetation, number of dif-ferent habitat types)

• presence of negative impact factors (e.g. road crossings and livestock)

Scores are averaged for each element and over-all, resulting in a single score between 1 and 10. The SVAP is typically sensitive enough to de-tect amelioration in stream and riparian condi-tions due to improved management practices at the scale of an individual large farm or group of smallholder farms.

The aquatic invertebrate assessment relies upon stream macroinvertebrates as a sensitive and well-tested proxy for a range of harder-to-measure water quality and stream habitat vari-ables. This indicator rates any given stream segment based on the abundance and diversity of pollution-tolerant versus pollution-sensitive macroinvertebrate families.

The water quality monitoring suite can be applied in its entirety if resources allow and the study or project design warrants it; in other cases, it can be streamlined to focus on key results or risks of interest. Based on recent testing and calibration of these meth-ods on cocoa farms in Ghana and coffee farms in Costa Rica, a set of guidelines for project managers and researchers is in development. These guidelines will select the most impor-tant components of the methodology to apply in any given context to maximize data quality while minimizing data collection costs and training requirements. Preliminary data from the water protocol comparisons in Ghana sug-

Crops and forest

interspersed in a diversified

rural landscape

JEFFREy HAyWARD

13

gest that some (but not all) low-tech, low-cost methods are acceptable substitutes for more costly probe-based methods on cocoa farms. In particular, water clarity tubes can replace turbidity probes; the float method can replace velocity meters; and the SVAP method corre-lates well with stream health as measured by macroinvertebrate assemblages.

Carbon and Greenhouse Gases

There is growing interest in monitoring the greenhouse gas “footprint” of supply chains for the purpose of corporate environmental reporting, consumer-facing marketing claims (e.g., “carbon neutral”), and possible participa-tion of farming communities in carbon markets. The scope for our carbon-monitoring work is “up to the farm gate”—that is, including the effects of on-farm management systems, land use, input use, and any on-farm processing, but not any subsequent processing, transport, refrigeration, etc. Since off-farm carbon emis-sions are generally not addressed by the SAN standard or the Rainforest Alliance, they are not a focus of our monitoring efforts but can be readily monitored through other systems.

The net on-farm greenhouse gas balance may be considered as the sum of: 1) greenhouse gas emissions (e.g., from fossil fuel use, emissions embedded in chemical inputs, and methane emissions) and 2) the net amount of carbon sequestered or emitted from land-use change. Most agricultural greenhouse-gas account-ing tools have focused on the first set of fac-tors. However, because the SAN standard is designed to increase on-farm vegetation (e.g., natural ecosystem conservation and tree can-opies for shade-tolerant crops), the second category is also important to consider in the context of the SAN standard. With this in mind, our greenhouse gas assessment tool is based on the Cool Farm Tool—an industry-leading method for agricultural greenhouse gas ac-counting—but adapts this tool to account for changes in aboveground carbon stocks associ-ated with on-farm vegetation cover. We do this by developing an inventory of non-crop vegeta-

tion patches on the farm (e.g., riparian forests and patches of high value ecosystems where cultivation is prohibited) based on the ecosys-tem type and successional state of each patch. For each patch, we apply appropriate carbon accumulation coefficients, based on prior stud-ies, to estimate annual aboveground carbon sequestration. These sequestration values are summed across all non-cropped patches on the farm (or within the land holdings of a farmer group). Any net carbon sequestration is then subtracted from the on-farm emissions esti-mate to arrive at the annual net on-farm green-house gas balance.

Farm Performance Monitoring Tool

As mentioned in the introduction, the adoption of more sustainable practices at the farm level is usually the first step toward the delivery of conservation benefits at all scales. By tracking changes in farmer practices over time, we can evaluate whether training and certification are delivering this important first step in the re-sults chain. In the absence of more elaborate (and expensive) monitoring of conservation outcomes, practice adoption may, in certain circumstances, also provide a suitable and relatively sensitive proxy indicator of whether farming systems are becoming more conserva-tion-friendly.

Coffeecherries

WILLIAM CRoSSE

14

case example Applying the Monitoring Framework to Chart Transitions in an Indonesian Cocoa-Producing Landscape The Rainforest Alliance’s environmental assessment framework can be considered as a “toolbox” from which the right methods are selected to meet the evaluation needs in any given landscape. In Bantaeng regency, at the southern tip of Indonesia’s Sulawesi Island, the Rainforest Alliance is working with cocoa farmers to increase crop productivity while improving the conservation value of the landscape. The landscape abuts the Gunung Lampobatang Protection Forest—the last remaining home of an endangered bird species—and furnishes clean water to downstream rice farms and communities. As cocoa demand in Indonesia grows, landscapes such as Bantaeng could be managed for increased production, thereby staving off pressures to convert cocoa farms to other (less conservation-friendly) crops and reducing the need to convert forest elsewhere for new cocoa farming. However, effective training and certification is needed to help overcome prior barriers to productive, conservation-friendly cocoa farming.

In this context, the project team identified three monitoring and evaluation needs to ensure that training and certification would meet these objectives, and that results could be rigorously documented. First, a thorough baseline assessment and risk analysis grounded in farmer reality was needed to identify the greatest barriers to sustainable production as well as farmer practices that did not comply with the SAN standard. To meet this need, we conducted community mapping with farmer groups and then applied the Farm Performance Monitor-ing Tool on a sample of 140 farms that would undergo training and certification.

Second, success from a conservation standpoint would hinge partly on increasing the conservation value of cocoa farms and adjacent areas. This would be reflected through changes in the quantity and diversity of non-crop vegetation in the landscape, with associated benefits for native wildlife, water quality, and erosion control. We applied the Natural Ecosystem Assessment (NEA) on-farm module for the same 140 farms to assess existing vegetation conditions, as well as the NEA landscape change assessment module (land cover mapping) to understand the broader context of habitats and wildlife corridors in the landscape.

Finally, the entire project approach—and the potential of more productive cocoa farming in Bantaeng to allevi-ate pressure on standing forests elsewhere in Indonesia—was predicated on increasing the productivity and profitability of cocoa farming. To this end, we assessed existing cocoa production and limitations through the Farm Performance Monitoring Tool as well as direct observation of cocoa plots (e.g., tree age, health, and yield estimated through pod counts) as part of the NEA on-farm module.

All of these methods were carried out by a team of local university students and recent graduates trained by Rainforest Alliance staff in use of the methods. Baseline assessments will be followed up 2-3 years later with repeat assessments, following training and certification. The use of standardized, well-tested field methods will ensure that baseline data collected in Bantaeng are comparable to data collected during other time periods and even at other sites, to carry out broader comparative or multi-site analysis.

We use a questionnaire called the Farm Perfor-mance Monitoring Tool as the basis for struc-tured interviews with farmers to understand farm management practices during the base-line (pre-training) period. This tool is imple-mented by local researchers, students, or para-professionals trained in the use of the method. Information generated through the tool is help-ful in designing effective training programs to address key barriers to conservation-friendly yet productive and profitable agriculture. The tool is also used to conduct follow-up assess-ments to track changes in practices over time. The tool collects data on three overall aspects of the farm and its management:

1. Basic information about the farm and farm-ing household, including household size and composition, crops, income sources, sales and recent investments

2. Information on each of the individual pro-duction plots within the farm, including the size, land use, types and management of crops

3. Information on the farm’s adherence to the requirements of the SAN standard, including key environmental management parameters such as management of natural ecosystems, shade tree management, soil conservation, agrochemical usage, pest and disease man-agement, hunting, human-wildlife conflict and record keeping

15

Although from a monitoring standpoint, a key goal is to assess the level of adoption of various best management practices included in the SAN standard, the Farm Performance Monitoring Tool serves an additional function of engaging the farmer in a conversation about how specific challenges (such as soil fertility management or maintenance of an appropriate shade-tree canopy) are being addressed on the farm. Such conversations can reveal important opportuni-ties or barriers to more productive and conser-vation-friendly agriculture—barriers that may need to be addressed through training, access to inputs or finance, or other mechanisms.

Community-Based Species Monitoring

One requirement within the Wildlife Conserva-tion principle of the SAN standard is the devel-opment of a farm species list, which is intended to raise awareness about local wildlife and pro-mote its conservation. The community-based species monitoring approach provides a tool to help these lists deliver greater value for farmers and for species monitoring. We begin by compiling a photographic key of locally oc-curring species of particular conservation or management concern. These include threat-ened species on the IUCN Red List and other species of local conservation concern, as well as pest species or invasive species that pose economic or safety risks to farmers. The pho-tographic key is used in the context of farmer interviews or community mapping to invite farmers to identify which of the animals they have observed and what their relationship is to each species (e.g., hunt it for food, appreciate it for its beauty, ignore it, or use deterrents to avoid crop-raiding). In this way, the accuracy of species-monitoring data is greatly improved, while farmers and trainers are able to identify the most critical risks and needs for avoiding problems such as human-wildlife conflict and illegal hunting. This tool is still in the pilot test-ing phase at sites in Indonesia and Ghana.

Next Steps for Charting Transitions to Conservation-Friendly Agriculture

With enough time and resources, monitoring environmental change is straightforward enough; scientists have been perfecting such methods for decades. The challenge that the Rainforest Alliance and its partners face is how

This Moor ma-caque, endemic to the island of Sulawesi and listed on the IUCN red list of endangered species, lives in cocoa-produc-ing landscapes where the Rainforest Alli-ance is working to promote sus-tainable land management practices.

THE RAINFoREST ALLIANCE

to monitor change and evaluate performance in the context of highly dynamic agricultural landscapes, diversified small-scale farms, and exacting supply chains that demand useful data at low cost. The system described here is our answer to this challenge. With a set of test-ed tools now in place, opportunities abound for collecting and using salient environmental data to guide farmer training, supply chain risk management, reporting on key perfor-mance indicators, and management of farming areas as parts of broader conservation strate-gies. We welcome new collaborations with farmer groups, traders, food companies, rural communities, students and researchers to apply these tools in the service of promoting productive, conservation-friendly agricultural in critical landscapes throughout the tropics.

16

233 broadway, 28th floornew york, ny 10279-2899T: +1.212.677.1900 f: +1.212.677.2187www.rainforest-alliance.org